One plugin, the whole stack

A Grafana app that wires your Prometheus, Loki, Tempo and cost data into curated Kubernetes views — no dashboard-hunting, no glue.

Live topology

Cluster status, namespace → workload → pod maps and node health, straight from the Kubernetes API via a read-only token.

Curated dashboards

Cluster overview, namespace, workload health, storage, control-plane and networking — computed inline from kube-state-metrics, node-exporter and cAdvisor.

Logs & traces inline

Click a pod to read its Loki logs in a drawer; explore Tempo traces and the service graph — built on Grafana Scenes.

Cost & right-sizing

Estimate cost from requests (or OpenCost), surface waste, and find over- and under-provisioned workloads.

RED & SLOs for any stack

Rate / errors / latency / error-budget burn rate from Istio, Linkerd, Cilium, NGINX, Traefik, Envoy, Kong and more — pick or auto-detect your stack.

Why KubeMeridian

Most Kubernetes monitoring setups are a pile of imported dashboards that drift, half-wired datasources, and a RED dashboard that only works for one ingress. KubeMeridian is the opposite: a single Grafana app where you point at your existing observability backends and immediately get a coherent, drill-downable view of every cluster.

Frontend-only & open source. No agent, no backend binary, Apache-2.0. It reads your Kubernetes API (read-only) and queries the datasources you already run.

Stack-agnostic. Metrics from any Prometheus-compatible store; logs from Loki; traces from Tempo; RED from whichever mesh/ingress you run (or a custom mapping).

Batteries included. 12+ bundled dashboards and a curated Prometheus alert pack ship with the plugin.

TIP

New here? Start with Install, then Connect your stack — that's the whole setup.

KubeMeridianKubernetes observability for Grafana